Alkyl Ketene Dimer (AKD) sizing – a review

Alkyl Ketene Dimer (AKD) sizing – a review 

Tom Lindström and Per Tomas Larsson, STFI-Packforsk AB, Stockholm, Sweden 

KEYWORDS: Alkyl ketene dimer, Internal sizing, 

Hydrophobicity, Retention, Spreading Reaction, Mechanisms, 

Paper, Board, Manufacture, Review 

SUMMARY: Over the years, there have been great efforts to 

try to develop cellulose reactive sizing agents. The assumption 

in these developments have been that the covalent linkage 

allows permanent attachment of hydrophobic groups in a highly 

oriented state, which makes sizing possible at very low levels of 

added chemical. The main requirement of the molecule is that it 

should have a balance between the reactivity towards water, 

because of the necessity of making stable emulsions or dispersions, 

and its reactivity towards cellulose. These assumptions 

are to some extent mutually exclusive and a compromise must 

be sought. Although, many different types have been tried out 

over the years the most important sizes used are the Alkyl 

Ketene Dimers (AKD) and the Alkenyl Succinic Anhydrides 

(ASA). These sizing agents are at the opposite in terms of stability 

of hydrolysis and reactivity towards cellulose, where AKDs 

are the least reactive species and fairly stable towards hydrolysis, 

whereas ASAs are very reactive towards cellulose, but also 

sensitive to hydrolysis. The mechanism of action is fairly well 

known for AKD, but less known for ASA and AKD-sizing can 

be regarded as a pretty mature field from a scientific point of 

view. The aim of this contribution is to summarize the fundamental 

features of AKD-sizing and discuss and highlight the 

most important aspects for the practical papermaker. 

Over the years there have been many reviews (e.g. (Dumas 

1975; Reynolds 1989; Eklund and Lindström 1991; Hodgson 

1994; Roberts 1997; Hubbe 2006)) in the field of AKD-sizing, 

but there have been extensive recent research activities over the 

past 10 years and there is a need for a comprehension of these 

research activities. 

ADDRESS OF THE AUTHORS: Tom Lindström 

(tom.lindstrom@stfi.se) and Per Tomas Larsson 

(tomas.larsson@stfi.se): STFI-Packforsk AB, Box 5604, 

SE-114 86 Stockholm, Sweden. 

Corresponding author: Tom Lindström 

Basic Chemical Features of AKD 

Alkyl ketene dimers were the results of direct development 

efforts in the late 40´s (Downey 1949). These 

investigations demonstrated that the parent molecule, the 

diketene could derivatize hydroxyl groups and in particular 

those of cellulose. The strained lactone ring in 

ketene dimers can react both with cellulose and water 

forming either the β-keto ester or the β-keto acid, which 

spontaneously decarboxylates to the corresponding ketone 

as shown in Fig 1. The ketone is incapable of reacting 

with cellulose. The balance of reaction with cellulose and 

hydrolysis is subtle, as with all reactive sizes, but the 

reaction is favoured and, hence, AKD can be used as a 

sizing agent under commercial papermaking conditions. 

The nucleophilic reaction with cellulose can also be accelerated 

with various so-called promoters, which will be 

discussed below. 

202 Nordic Pulp and Paper Research Journal Vol 23 no. 2/2008 

Fig 1. Alkyl ketene dimers can react either with cellulose forming the β-keto ester 

or with water forming the β-keto acid, which spontaneously decarboxylates forming 

the corresponding ketone. 

Commercial AKDs are prepared from long fatty acids via 

their acid chlorides, which then dimerize to the corresponding 

alkyl ketene dimer (Fig 2). 

Fig 2. Formation of alkyl ketene dimers from the corresponding fatty acid chlorides. 

The linear saturated AKDs are waxy substances, water 

insoluble solids with melting points around 50°C, being 

mixtures of C-14 to C-18 fatty acids when manufactured 

from commercially available fatty acids. C-16 is the predominant 

fatty acid in the most used formulations. The 

efficiency increases with carbon chain length from C-8 

and levels off at C-20 (Brungardt and Varnell 2005). 

It is also known that small amounts (usually less than 

10% (Hardell and Woodbury 2002) of AKD-oligomers 

are present in AKD (Bottorff 1993; Bottorff 1994; 

Asakura et al. 2006a) and that these oligomers give 

relatively poor sizing. 

Alkenylsubstituted AKDs (emulsions = liquid in liquid 

phase) are also commercially available and they have 

been found to give better runnability on high speed 

handling equipment than paper sized with suspensions 

(solid phase in a liquid phase) (Brungardt and Gast 

1996), but they are slightly inferior to alkyl ketene dimers 

(Isogai and Asakura 2001a, 2001b). 

Emulsification 

AKDs are dispersed using high-pressure homogenizers at 

elevated temperatures. The most frequently used 

stabilizers are cationic starches in conjunction with 

lignosulfonates/naphthalene sulfonic acids. Waxy maize 

starches with no propensity to retrogradation are the 

preferred choice of starch. Most commercially used 

dispersions (note: both emulsions and suspensions are dispersions) 

are therefore amphoteric (Isogai 1997b), which 

is important from a retention point of view (see below).

It is important to avoid surface active substances in the 

dispersion formulation, because they may interfere with 

sizing (see below). The dispersions are usually made 

slightly cationic in order for them to have a natural 

substantivity to negatively charged fibres, but anionic 

dispersions are also used in commercial practice. In order 

to avoid hydrolysis of the AKD, the pH is kept around 3 

in the formulations. 

There have been some recent studies on the stability of 

AKD-dispersions. It has been found that the stability 

increases by the presence of AKD-oligomers (Asakura et 

al. 2006a) and that fatty acid anhydrides (byproduct in 

the AKD-dispersion) decrease the heat shock stability 

(Asakura et al. 2006b). The effect of various colloidal 

substances present in process waters have also been investigated 

(Mattsson et al. 2001). AKD dispersions basically 

behave as electrostatically stabilized colloids. 

Consecutive events of AKD-sizing 

The consecutive events in AKD-sizing are (Lindström 

and Söderberg 1986a): 

· Retaining the AKD-size using appropriate retention 

strategies for the size 

· Spreading/size migration to a monolayer 

· Chemical reaction of the size with the cellulosic fibres 

The retention mechanism is, in theory, heterocoagulation, 

where cationic size particles are attached to the negatively 

charged fibres. This is expected to give a good 

distribution of the dispersion particles, but practice shows 

that size distribution is not critical, because of extensive 

spreading on the fibre surfaces. More important is that 

effective retention aids must be used for the purpose. 

Heterocoagulation, cannot be used to retain AKD-particles 

under high fibre-fibre shear conditions. A high 

single pass retention is important, because recirculated 

size is hydrolyzed in the white water of a paper machine. 

When the size particles have been deposited on the 

fibres, the reaction with cellulose is insignificant because 

very few molecules are in molecular contact with cellulose. 

Because of the high surface tension of water, no spreading 

can take place until the water has been removed 

and the size particles are in direct contact with air. 

Hence, an air-AKD surface must be formed before 

spreading can take place and this takes place during 

drying at a solids content exceeding 60%. The spreading 

continues until a monomolecular layer has been formed. 

This layer then reacts with the hydroxyl groups of 

cellulose. The reaction is slow at low pH-values and, in 

practice; AKD cannot be used except in the neutral or 

slightly alkaline pH-range. Moreover, sizing accelerators 

(most commonly HCO - 

3) are almost always required for 

effective development of sizing. 

Retention of AKD 

Cationic AKD-particles are theoretically retained by a 

heterocoagulation deposition mechanism onto the negatively 

charged fibres, but studies (Johansson and 

Lindström 2004b) have shown that the electrostatic 

attraction is screened by electrolyte concentrations 

commonly present in process waters, so retention aids 

must always be used in commercial practice. Secondly, 

the deposition is a highly dynamic process, where AKDparticles 

are rapidly deposited, after which they are 

sheared off the fibre surfaces (Lindström and Söderberg 

1986c; Champ and Ettl 2004). 

The surface charge of fibres is important for the selfretention 

(retention of AKD without retention aid 

addition) of AKD, as self-retention is a heterocoagulation 

mechanism (Isogai et al. 1997; Lindström and Glad- 

Nordmark 2007a). There is an optimum surface charge 

density of fibres and an optimum electrolyte concentration 

for maximum AKD-retention (Lindström and Glad- 

Nordmark 2007a). These studies are, however, laboratory 

studies, and in commercial practice self-retention is most 

likely negligible due to high electrolyte concentrations 

and the high shear to which the fibre suspension is 

subjected. 

The retention of AKD by cationic polyelectrolytes has 

been studied by several other groups (Esser and Ettl 

1997; Hasegawa et al. 1997; Isogai 1997a, 1997b). It is 

not self-evident that it should be possible to retain 

cationic AKD-dispersions using cationic polyelectrolytes. 

Due to the fact that AKD-dispersions usually are 

dispersed using a combination of lignosulfonates and 

cationic starch, the net cationically charged dispersion 

particles have in fact an amphoteric nature. Hence 

cationic polyelectrolytes can interact with the negative 

sites on the dispersion particles and retain the AKD-particles. 

As shown in Fig 3, anionic AKD-particles are easier 

to retain by cationic polyelectrolytes than cationic AKDparticles 

(Johansson and Lindström 2004b). The 

appropriate choice of an efficient retention aid system is 

crucial, because AKD will be subject to hydrolysis when 

circulated in the short circulation of a paper mill, as will 

be discussed below. 

The use of hydrophobic retention aids has also been 

practised (Riebeling et al. 1996, 1999) to treat filler so 

that the interaction between filler particles and AKD 

could be minimized. Obviously AKD cannot react with 

fillers and CaCO 3-based fillers promote hydrolysis (see 

below). 

Pre-flocculation of AKD leads to higher AKD-retention 

(Mattsson et al. 2002), which in itself is beneficial to 

sizing. Agglomeration is not critical to sizing (Johansson 

and Lindström 2004b); neither is the particle size 

(Petander et al. 1998) because AKD spreads over the 

fibre surfaces anyhow. 

In the practical use of AKD, the point of addition is of 

course dependent on the wet end system and the addition 

order of other chemical adjuvants. As a rule of thumb it is 

advantageous to add AKD to the high consistency stock 

only a short time before dilution takes place in the short 

circulation. The AKD-particles are deposited at a rapid 

rate at high consistency but are subsequently sheared off 

the fibre surfaces by the other fibres in a stirred pulp 

suspension (Lindström and Söderberg 1986a; Champ and 

Ettl 2004). Both high shear and long contact times are 

known to reduce AKD retention. As a rule, both AKD, 

Nordic Pulp and Paper Research Journal Vol 23 no. 2/2008 203

Fig 3. a) Retention of cationic AKD-particles onto bleached kraft pulps using various cationic polyelectrolytes. b) Retention of anionic AKD-particles onto bleached kraft pulp 

using various cationic polyelectrolytes (Johansson and Lindström 2003b). 

fines and fillers follow the general retention level and 

there is little selective retention of certain types of dispersed 

material. Anionic dissolved substances in the stock, 

such as hemicelluloses and lignin residues, are generally 

detrimental to size retention (Lindström and Söderberg 

1986c). 

Spreading/size migration 

The distribution of AKD-size on the fibres occurs in the 

drying section as discussed above. AKD readily spreads 

on cellulose, because the cellulose surface is a high-energy 

surface. The free energy of spreading, ∆G s of AKD on 

a cellulose surface can be written: 

∆G s = γ(cellulose/AKD) + γ(AKD) - γ (cellulose) [1] 

The surface free energy of AKD is 27 mJ/m 2 (Garnier 

and Godbout 2000) and the surface free energy of dry 

cellulose has been determined to about 57 mJ/m 2 . 

(Lundqvist and Ödberg 1997; Luner and Oh 2001). If 

γ(cellulose/AKD) is small, the conditions for spreading, 

∆G s < 0, are fulfilled. For an AKD particle trapped in 

between a fibre-fibre bond the free energy of spreading, 

∆G s = 2γ(cellulose/AKD), which is a positive quantity, 

because it is associated with the cleavage of a high 

energy surface. Hence, AKD-particles trapped in between 

fibre-fibre bonds cannot spread and cannot react with 

cellulose. This phenomenon is the hypothesis for the 

limited reaction with cellulose (Lindström and O´Brian 

1986b). Spreading has been manifested by several investigators 

(Roberts and Garner 1985; Roberts et al. 1985; 

Ödberg et al. 1987; Seppänen et al. 2000; Horn 2001; 

Shchukarev et al. 2003). The spreading of AKD on 

cellulose should, however, not be associated with the 

common hydrodynamic phenomena of spreading 

(Cazabat 1989), which is a very rapid process. Instead, it 

has been suggested that spreading takes place by the 

surface diffusion of an autophobic monolayer of AKD on 

cellulose (Seppänen et al. 2000), which is a slower 

phenomena than hydrodynamic spreading. The apparent 

surface diffusion coefficient of AKD on cellulose have 

been calculated to around 10 -11 m 2 /s at 50-80°C 


(Seppänen et al. 2000; Shchukarev et al. 2003). As AKDsizing 

particles typically have the dimension of the order 

of a micrometer, such a droplet would on a cellulose 

surface spread within 10 sec, using this diffusion 

coefficient. The time of reaction is typically of the order 

of at least 5 minutes, hence spreading/surface diffusion is 

not the rate-determining step in sizing. 

More recent investigations have, however, challenged 

the traditional spreading view. Thus, Garnier et al 

(Garnier et al. 1998, 1999; Garnier and Godbout 2000) 

find that AKD only wets, but not spreads, on cellulose 

and claim vapour phase sizing as a sizing mechanism for 

AKD. Later investigations show vapour phase type of 

sizing with ASA but not with AKD at drying temperatures 

below 100°C. (Yu and Garnier 2002). A capillary 

wicking mechanism is also suggested by Garnier and 

Godbout (2000). One possibility is that the reservoir in 

the de Gennes type of experiments conducted by Garnier 

was simply too small. In this type of experiment a drop of 

the sizing agent is applied to a thread (cellulosic) and the 

contact angle is observed. A pre-requisite is that there is a 

sufficient surface area available for spreading to occur, 

otherwise spreading will stop when the available surface 

area is saturated with the monomolecular layer of the 

sizing agent, spreading stops and a finite contact angle is 

observed. 

Shen et al. also in a number of papers (Shen et al. 

2001a, 2001b; Shen and Parker 2003) claim that the 

classical view has been proven wrong and advocate 

mechanisms along the lines of Garnier and co-workers. 

These authors also claim there is no reaction between 

cellulose and AKD. Later investigations from this group 

(Hutton and Chen 2004), however, also show vapour 

phase sizing to be an insignificant phenomenon in 

practical papermaking sizing. The group has now (Shen 

and Parker 2003; Shen et al. 2005) adopted the autophobic 

precursor mechanism suggested by Seppänen et 

al. (2000). 

In our laboratory, a simple experiment was conducted 

to show that spreading does not take place in the gasphase 

and that spreading easily takes place and over 

macroscopic dimensions. Basically, two sets of experi-

ments were conducted in additon to a reference sheet, 

# 1 (Table 1). In the first experiment (# 2, Table 1) a sized 

sheet was couched (after wet-pressing) between two 

unsized sheets. This pack of sheets was then dried at 

90ºC and the size retention and extent of reaction was 

determined using radioactive labelled AKD as described 

in our previous publications on AKD, e.g., Lindström and 

Söderberg (1986a). After drying the stack of sheet was 

simply delaminated and the total and reacted amount of 

AKD could be determined. In the second experiment 

(# 3, Table 1) a thin polytetrafluoroethylene (PTFE) wire 

(allowing gas-phase transfer through the wire) was 

placed between the sheets before drying. 

These experiments showed two things: AKD could 

spread across the whole stack of sheets, although the 

mid-sheet still had the highest amount of AKD. In the 

experiment using the PTFE wire, no transfer between the 

sheets took place and consequently there was no gasphase 

spreading across the pile of sheets. 

Table 1.Results from AKD-transfer experiments. AKD-sized sheets (2.2 and 3.2) 

were stacked between unsized sheets (sheets: 2.1, 2.3 and 3.1, 3.3) after wet 

pressing and subsequently dried at 90ºC (series # 2) for 10 min. After drying, the 

sheets were post cured at 110ºC for 10 min. The reference sheet was dried at 

90ºC for 10 min and then at 110ºC for 10 min. Experiment # 3 was conducted as 

2.1-2.3 but a thin polytetrafluoroethylene wire was inserted between the sheets 

before couching to prevent size migration but allow for possible gas phase transfer 

of AKD. 0.1% C-PAM (D.S. = 20 mole % cationic groups) was used as a 

retention aid in the experiments. n.d. = not determined. Data not published previously. 

Sheet Content AKD AKD AKD 

in sample retention reacted amount 

Exp. #/Sheet type mg/g % mg/g % 

1.1 AKD (1.5 mg/g) 1.15 76.7 0.46 40.0 

(Reference) +C-PAM 

2.1 no AKD 0.16 0.07 43.8 

2.2 AKD (1.5 mg/g) 0.88 80.0 0.38 43.2 

+C-PAM 

2.3 no AKD 0.16 0.06 37.5 

3.1 no AKD [AKD] and the equation reads: 

dx 

dt 

( ) 

= K [AKD] = K [AKD] = K a− x 

[3] 

av 

1 2 

where a is the maximum amount of reacted AKD and 

consequently [AKD] av, the available amount for reaction 

should be = a – x. Integration from t = 0 to t leads to: 

ln a 

a x Kt = [4] 

− 

Thus, if the quantity ln(a/(a – x)) is plotted versus time, 

straight lines should be obtained and the reaction follows 

what could be called a pseudo first order reaction. This is 

illustrated in Fig 4a, which shows such graphs for the 

reaction of AKD onto a bleached kraft pulp at different 

pH-values. It is also shown that the straight lines intersect 

the x-axis at a common point in time, independent of pH. 

This is the time interval required for drying of the sheet 

to a sufficiently high solids content for the AKD to begin 

spreading. The spreading reaction as such is much faster 

than the drying stage, which is independent of the pHvalue 

as the rate of drying is independent on pH for this 

type of pulp. If the same experiment then is performed at 

different pH-values, the corresponding values between 

the reaction rate constant K at different pH-values and 

temperatures can be obtained and the Arrhenius activation 

energies for the AKD/cellulose reaction can be 

calculated from the graphs in Fig 4b. 

The calculated activation energies for the reaction 

between cellulose and AKD can be calculated from 

Fig 4b to be 72 kJ/mole at pH 4 to 46 kJ/mole at pH 10 

(Lindström and O´Brian 1986b). As expected the activa- 

Fig 4. (a) The quantity In a/(a-x) versus reaction time, t, where a is the maximum 

reacted AKD and x is the reacted amount of AKD at different pH-values. (b) The 

reaction rate constant K, from Eq 2, versus 1/T, where T is the absolute temperature. 

(Bleached softwood kraft pulp). (Lindström and O´Brian 1986b). 


tion energy decreases for the nucleophilic reaction the 

higher the pH. 

Sizing accelerators 

The reaction between AKD and cellulose is, however 

slow and sizing accelerators are invariably used in commercial 

operations, whereby the reaction rate easily can 

be increased 20 times. The most important sizing accelerators 

are: 

· HCO-3 

· Basic polymers with amine groups 

The HCO - 

3-ion has a unique ability to catalyse the reaction 

between AKD and cellulose. HCO - 

3- -ions are often 

inherently present in natural systems, e.g. when CaCO3 is 

used as filler, but it is also a general practice to add 

NaHCO3 to increase the alkalinity of the stock. 

Fig 5a shows how the reaction of AKD in the presence 

of NaHCO3 is catalysed. The acceleration has been 

quantified using the above reaction rate expression. A 

further analysis of the so obtained reaction rate constant 

reveals that the reaction rate is proportional also to 

[HCO - 

3]. Hence, the reaction follows the equation: 

dx 

dt K 

- 

= [cellulose] ⋅[AKD] ⋅[HCO 

] 

0 3 

This equation clearly suggests that there is a trimolecular 

reaction between cellulose, AKD and HCO 3 taking place. 

The suggested mechanism of catalysis is given in Fig 5b. 

Polymeric amines (e.g. PAMAM-EPI resins) having 

amino groups with a free electron pair are classic sizing 

accelerators for AKD (Lindström and Söderberg 1986d; 

Thorn et al. 1993; Cooper et al. 1995). Several different 

Fig 5. (a) The quantity ln a/(a-x) versus reaction time, t, where a is the maximum reacted AKD and x is the reacted 

amount of AKD at different bulk concentrations of NaHCO 3. (b) A suggested mechanism of catalysis with NaHCO 3. 

(Lindström and Söderberg 1986d). 


[5] 

Fig 6. Synergistic effects between HCO - 

3and PAMAM-EPI resin on the reaction 

rate constant, K, when simultaneously used in AKD-sizing. ❍ = sizing in deionized 

water at different pH-values. ●– = sizing in the presence of 0.1% PAMAM in deionized 

water (Lindström and Söderberg 1986a). 

types of condensation polymers have been investigated 

over the years and occur in commercial formulations. The 

polymeric sizing accelerators are often either added to the 

AKD-dispersion (”rapid curing dispersion”) or used separately 

as combined accelerator and retention aid. Moreover 

the effects of HCO - 

3 are synergistic, as shown in Fig 6. 

AKD-hydrolysis 

AKD can also react with water forming the β-keto acid, 

which spontaneously decarboxylates forming the corresponding 

ketone, as shown in Fig 1. AKD is, however, 

stable at room temperature at acidic pH-values, allowing 

storage at the time scale of months. 

It is well known that AKD can be 

the subject to alkaline hydrolysis. It 

is known that CaCO 3 may induce 

hydrolysis, particularly precipitated 

calcium carbonates (PCC) having 

higher pH-values due to residual 

alkali (Colasurdo and Thorn 1992; 

Novak and Rende 1993; Bottorff 

1994; Jiang and Deng 2000). As 

AKD is strongly adsorbed onto PCC 

and gronud calcium carbonate 

(GCC), it is advantageous to preadsorb 

cationic polymers onto carbonate 

fillers in order to block AKDdeposition 

(Esser and Ettl 1997). 

There have, however, been few systematic 

and quantitative investigations 

in this field. 

In a recent investigation from this 

lab, the hydrolysis was studied using 

14 C-labelled AKD (Lindström and 

Glad-Nordmark 2007b). It was 

found that NaHCO 3 catalyzed the 

hydrolysis reaction, as shown in

Fig 7a. The mechanism is probably 

analogous to the mechanism of the 

catalysis of the cellulose-AKD 

reaction shown in Fig 7b, because 

analysis of the reaction kinetics of the 

reaction pointed to a trimolecular 

reaction mechanism. In this investigation 

it was also found that divalent 

metal ions (Ca 2+ , Mg 2+ , Ba 2+ ) catalyzed 

the hydrolysis reaction (see Fig 8a), 

whereas the anion or monovalent electrolytes 

had no effects. It is most likely 

that the divalent cation functions 

as a Lewis type of catalyst (Schinzer 

1986), the mechanism of which is 

shown in Fig 8b. Hydrolysis only 

takes place at high temperatures and 

during standard laboratory forming 

and drying conditions hydrolysis is 

insignificant. 

The hydrolysis product of AKD, the 

ketone in Fig 1, has a slight positive 

effect on sizing, provided there is already 

some reacted AKD present in the 

paper (Lindström and Söderberg 

1986a). 

Amount of AKD required for sizing 

The required amount of AKD for 

sizing for a given pulp depends on a number of factors 

and is also linked to a number of wet-end factors. Critical 

is the retention of the size and the extent of reaction 

together with the nature of the pulp furnish together with 

the structure of the sheet. 

Retention is critical, as recirculated size can be subject 

to hydrolysis. The extent of reaction depends on the 

drying conditions together with the presence of size 

accelerators. The extent of reaction for AKD-sizes is 

dependent on the fraction of fibre surface exposed to the 

air phase, because it is only size on free surfaces, which 

can spread and potentially react. AKD spreads all over 

the sheet and sizing with AKD is not dependent on size 

agglomeration, which is critical for instance with 

soap/alum sizing. 

The extent of reaction can be quite high under ideal 

laboratory conditions, but under practical mill conditions 

it is often in the range between 15-40%. 

In an early publication (Lindström and Söderberg 1986a) 

the amount required for sizing was investigated for 

various extractive free pulps. Defining the onset of full 

sizing as Cobb 60 = 25 g/m 2 , the required amount of AKD 

necessary for sizing is directly proportional to the BET 

surface area of the papers as shown in Fig 9. By using 

surface balance measurements to determine the collapse 

value of the monolayer the planar oriented monolayer 

surface area of AKD can be calculated to 24 Å 2 per molecule. 

It is important to emphasize that sizing is uniquely 

defined by the reacted amount of AKD (Lindström and 

Söderberg 1986a; Johansson and Lindström 2004b). 

Using this surface area it can be calculated from Fig 9 

Fig 7. (a) The effect of NaHCO 3 on AKD hydrolysis, (b) Suggested hydrolysis mechanism. 

Fig 8. (a) The effect of CaCl 2 on AKD hydrolysis, (b) Suggested hydrolysis mechanism by Ca 2+ (Lewis acid catalysis). 

(Lindström and Glad-Nordmark, 2007b). 

Fig 9.The required amount of reacted AKD, necessary to obtain Cobb 60 = 25 g/m 2 

for various pulps (extractives free). 

that it is only necessary to cover 4% of the total surface 

area for a given pulp in order to obtain sizing. In practice 

the sweeping action of the fatty acid chains will, of course, 

cover a much larger area. Ström and co-workers 

(Ström et al. 1992) determined the required surface coverage 

to 15% using ESCA. The electrons from carbon 

atoms emitted in the ESCA analysis partly come from 

carbon atoms underneath the fatty acid layer, so the surface 

coverage values should not be directly compared. 


Practical aspects and comparisons between different 

sizing agents 

In Table 2, some major aspects of different sizing agents 

have been compared. Rosin sizing is basically restricted 

to acidic pH-values and so both AKD and ASA are basically 

restricted to neutral/alkaline papermaking, although 

ASA may be used at slightly acidic pH-values (Roberts 

1997). Electrolytes are basically negative for all sizing 

agents because they interfere with retention aid use and 

decrease their affinity to fibre surfaces. Divalent metal 

ions are devastating for rosin sizing because they compete 

with aluminium species in the complexation process 

and for AKD sizing, because they catalyze the AKD 

hydrolysis reaction. Fines/fillers have a large surface area 

consuming the sizing agent. Fillers can generally not be 

sized because reactive sizes do not react with fillers and 

the aluminium resinate (rosin-aluminium complex) 

cannot be anchored to the filler. 

The hydrolysis product of ASA can complex with Ca 2+ - 

ions so it may in principle be able to use for slack sizing 

of calcium carbonates (Roberts 1997). 

Dissolved anionic substances are in almost all cases 

detrimental to size retention. The charged groups on the 

fibres are necessary for rosin sizing, and the higher the 

carboxyl group content, the easier is is to size the pulp 

with rosin. For AKD/ASA sizes, the charged groups are 

in general beneficial for retention processes. Acidic 

extractives may in principle be used as sizing agents in 

the presence of alum and non-ionic extractives contribute 

only slightly to sizing. Extractives of the fatty acid types 

are detrimental to AKD-sizing, because they interfere 

with retention and with the AKD-reaction (Lindström 

and Söderberg 1986c; Lidén and Tollander 2004; Åvitsland 

et al. 2006), but have been found not to interfere 

with spreading (Mattsson et al. 2003). Extractives have in 

general a slight positive effect on ASA-sizing because 

aluminium salts are used in conjunction with ASAsizing. 

The AKD-hydrolysis product has a slight positive 

effect on AKD-sizing, but is detrimental for ASA-sizing 

because the diacid is amphiphatic and will overturn in the 

presence of aqueous liquids in contact with the sized 

paper. Aluminium salts are, of course, necessary for rosin 

sizing, but interfere with AKD-sizing, if they contribute 

sufficient acidity to decrease the HCO - 

3 content of the 

water. AKD can also be used in conjunction with rosin 

sizing for liquid packaging (Walkden 1991), but there is 

no simple mechanistic explanation for this synergism. 

Table 2. Comparison between different sizing agents. 

Rosin AKD ASA 

pH 4.2-5.0 7-8.5 5-8.5 

Electrolytes - - - - - 

Fines/Fillers - - - 

Dissolved an. substances - - - - - - 

Fibre-COOH + + + + 

Extractives + - - (+) 

Hydrolysis products irrelevant (+) - - 

Aluminium sulfate + + + 

Stock temperature - - - - 

Lactic acid resistance - + - 


Stock temperature has a strongly negative effect, particularly 

on rosin soap sizing, but also on dispersion sizing. A 

higher temperature leads to aggregation of precipitated 

aluminum resinates and oxolation of aluminium species 

making them loose some of their cationic charge characteristics. 

For synthetic sizes a higher stock temperature 

leads to a higher rate of hydrolysis of non-retained size. 

Neither rosin sizes nor ASA-sizes can protect paper 

against liquids containing strongly coordinating species 

like lactic acid 

Acknowledgement: 

Veronica Sundling is acknowledged for editing the text. 

Literature 

Asakura, K., Iwamoto, M. and Isogai, A. (2006a): The effects of AKD oligomers 

present in AKD wax on dispersion stability and paper performance, Nord. Pulp 

Paper Res. J. 21(2), 245-252. 

Asakura, K., Iwamoto, M. and Isogai, A. (2006b): Influence of fatty acid anhydride 

components present in AKD wax on emulsion stability and paper sizing performance, 

APPITA, 59(4), 285-290. 

Bottorff, K.J. (1993): New insights into the AKD sizing mechanism, Nord. Pulp 

Paper Res. J. 8(1), 86-95. 

Bottorff, K.J. (1994): AKD sizing mechanism: A more definitive description, Tappi 

J. 77(4), 105-116. 

Brungardt, C.L. and Gast, J.C. (1996): Alkenyl-substituted sizing agents for precision 

converting grades of fine paper, Tappi Proc. 1996 Papermakers 

Conference, TAPPI Press, Atlanta, USA, 297-308. 

Brungardt, C.L. and Varnell, D.F. (2005): The effect of ketene dimer melting 

point on the rate of sizing development, In: Advances in Paper Science and 

Technology: 13th Fundamental Research Symposium, Cambridge, I´Anson, S.J. 

(ed.), Pulp and Paper Fundamental Research Society, Bury, UK, 193-209. 

Cazabat, A.M. (1989): The dynamics of wetting, Nord. Pulp Paper Res. J 4(2), 

146-154. 

Champ, S. and Ettl, R. (2004): The dynamics of alkylketene dimer (AKD) retention, 

J. Pulp Paper Sci. 30(2), 322-328. 

Colasurdo, A.R. and Thorn, I. (1992): The interaction of alkyl ketene dimer with 

other wet-end additives, Tappi J. 75(9), 205-211. 

Cooper, C., Dart, P., Nicholass, J. and Thorn, I. (1995): The role of polymers in 

AKD sizing, Paper Technol. (May), 30-34. 

Downey, W.F. (1949): Higher alkyl ketene dimer emulsion. US Pat. 2,627, 477. 

Dumas, R.W. (1975): An overview of cellulose reactive sizes, Tappi J. 64(1), 43. 

Eklund, D. and Lindström, T. (1991): Paper Chemistry – An introduction, DT 

Paper Science Publications, Grankulla, Finland. 

Esser, A. and Ettl, R. (1997): On the mechanism of sizing with alkyl ketene dimer 

(AKD): physico-chemical aspects of AKD retention and sizing efficiency, 

Fundamentals of Papermaking Materials: 11th Fund. Res. Symp. Cambridge, 

Fundamental Res. Comm. and Pira International, 997-1020. 

Garnier, G., Bertin, M. and Smrckova, M. (1999): Wetting dynamics of alkyl 

ketene dimer on cellulosic model surfaces, Langmuir, 15(22), 7863-7869. 

Garnier, G. and Godbout, L. (2000): Wetting behaviour of alkyl ketene dimer on 

cellulose and model surfaces, J. Pulp Paper Sci. 26(5), 194-199. 

Garnier, G., Wright, J., Godbout, L. and Yu, L. (1998): Wetting mechanism of 

alkyl ketene dimers on cellulose films, Coll. Surf. A: Phys. Chem. Eng. Asp. 145, 

153-165. 

Hardell, H.-L. and Woodbury, S. E. (2002): A new method for the analysis of AKD 

oligomers in papermaking systems, Nord. Pulp Paper Res. J. 17(3), 340-335. 

Hasegawa, M., Isogai, A. and Onabe, F. (1997): Alkaline sizing with alkylketene 

dimers in the presence of chitosan salts, J. Pulp Paper Sci. 23(11), J528-J531. 

Hodgson, K.T. (1994): A review of paper sizing using alkyl ketene dimer vs. alkenyl 

succinic anhydride, APPITA, 47(5), 402. 

Horn, D. (2001): Exploring the nanoworld of interfaces and their functions

during paper manufacturing and upgrading, Wochenbl. für Papierfab. 

129(23/24), 1589-1596. 

Hubbe, M. ,A. (2006): Paper´s resistance to wetting – A review of internal sizing 

chemicals and their effects, Bioresources, 2(1), 106-145. 

Hutton, B. and Chen, W. (2004): Sizing effects via AKD vaporisation, 58th Appita 

Annual Conf. and Exhibition, Canberra, ACT, Australia. 

Isogai, A. (1997a): Factors influencing on retention of alkyl ketene dimer, The 

Fundamentals of Papermaking, Trans. of the 11th Fund. Res. Symp. held at 

Cambridge, Pira International, 1047-1071. 

Isogai, A. (1997b): Effect of cationic polymer addition on retention of alkyl ketene 

dimer, J. Pulp Paper Sci. 23(6), J276-J281. 

Isogai, A. and Asakura, K. (2001a): Paper sizing by liquid – type of ketene 

dimers Part 1. Comparison with alkyl ketene dimer and oleic anhydride, Nord. 

Pulp Paper Res. J. 16(2), 103-107. 

Isogai, A. and Asakura, K. (2001b): Paper sizing by liquid-type ketene dimers. 

Part 2. Effects of drying methods, UV and air-blowing treatments on PCC-filled 

sheets, Nord. Pulp Paper Res. J. 16(2), 108-112. 

Isogai, A., Kitaoka, C. and Onabe, F. (1997): Effects of carboxyl groups in pulp 

on retention of alkyl ketene dimer, J. Pulp Paper Sci. 23(5), J217-J219. 

Jiang, H. and Deng, Y. (2000): The effects of inorganic salts and precipitated 

calcium carbonate filler on the hydrolysis kinetics of alkyl ketene dimer, J. Pulp 

Paper Sci. 26(6), 208-213. 

Johansson, J. and Lindström, T. (2004b): A study on AKD-size retention, reaction 

and sizing efficiency. Part 2: The effects of electrolytes, retention aids, shear 

forces and mode of addition on AKD-sizing using anionic and cationic AKD-dispersions, 

Nord. Pulp Paper Res. J. 19(3), 336-344. 

Lidén, J. and Tollander, M. (2004): Extractives in totally chlorine free bleached 

birch pulp and their effect on alkyl ketene dimers and alkeny succinic anhydride 

sizes, Nord. Pulp Paper Res. J. 19(4), 466-469. 

Lindström, T. and Glad-Nordmark, G. (2007a): A study of AKD-size retention, 

reaction and sizing efficiency. Part 3: The effects of fibre charge density and electrolyte 

concentration on size retention, Nord. Pulp Paper Res. J. 22(2), 161-166. 

Lindström, T. and Glad-Nordmark, G. (2007b): A study of AKD-size retention, 

reaction and sizing efficiency. Part 4: Mechanism of AKD-hydrolysis, Nord. Pulp 

Paper Res. J. 22(2), 167-171. 

Lindström, T. and O´Brian, H. (1986b): On the mechanism of sizing with alkyl 

ketene dimers.Part II. The kinetics of reaction between alkyl ketene dimers and 

cellulose, Nord. Pulp Paper Res. J. 1(1), 34-42. 

Lindström, T. and Söderberg, G. (1986a): On the mechanism of sizing with alkyl 

ketene dimers. Part I. Studies on the amount of alkyl ketene dimer required for 

sizing of different pulps, Nord. Pulp Paper Res. J. 1(1), 26-33. 

Lindström, T. and Söderberg, G. (1986c): On the mechanism of sizing with alkyl 

ketene dimers. Part III. The role of pH, electrolytes, retention aids, extractives, Calignosulfonates 

and mode of addition on alkyl ketene dimer retention, Nord. Pulp 

Paper Res. J, 1(2), 31-38. 

Lindström, T. and Söderberg, G. (1986d): On the mechanism of sizing with alkyl 

ketene dimers. Part IV. The effects of HCO 3-ions and polymeric reaction accelerators 

on the rate of reaction between alkyl ketene dimers and cellulose, Nord. Pulp 

Paper Res. J. 1(2), 39-45. 

Lundqvist, Å. and Ödberg, L. (1997): Surface energy characterization of surface 

modified cellulosic fibres by inverse gas chromatography, in The Fundamentals of 

Papermaking Materials. Trans of the 11th Fund. Res. Symp. Sept. 1997, held at 

Cambridge, UK, Pira International, Leatherhead, Surrey, UK, Vol 1, pp. 751-769. 

Luner, P.E. and Oh, E. (2001): Characterization of the surface free energy of cellulose 

ether films, Coll. Surf. A. Phys. Chem. Eng Asp. 181, 31-48. 

Mattsson, R., Lindström, D., Sterte, J. and Ödberg, L. (2003): Influence of abietic 

acid, betulinol, sodium oleate and tripalmitine on the migration of AKD in 

paper, J. Pulp Paper Sci. 29(8), 281-285. 

Mattsson, R., Sterte, J. and Ödberg, L. (2001): Colloidal stability of alkyl ketene 

dimer (AKD) dispersion: Influence of shear, electrolyte concentration, polyelectrolytes 

and surfactants, The Science of Papermaking, 12th Fundamental Res. 

Symp. Oxford, The Pulp and Paper Fundamental Res. Soc., 393-413. 

Mattsson, R., Sterte, J. and Ödberg, L. (2002): Sizing with pre-flocculated alkyl 

ketene dimer (AKD) dispersions, Nord. Pulp Paper Res. J. 17(3), 240-245. 

Nahm, S. (1986): Direct evidence for covalent bonding between ketene dimer 

sizing agents and cellulose, J. Wood Chem. Technol. 6(1), 89-112. 

Novak, R. and Rende, D. (1993): Size reversion in alkaline papermaking, Tappi J. 

76(8), 117-120. 

Petander, L., Ahlskog, T. and Juppo, A. J. (1998): Strategies to reduce AKD 

deposits on paper machines, Paperi Puu, 80(2), 100-103. 

Reynolds, W. F. (1989): The Sizing of Paper, Tappi Press, Atlanta, GA, USA. 

Riebeling, U., Jeurissen, H.F.M., De Clercq, A. and Prinz, M. (1996): Ein neues 

Masseleimungskonzept-Wirkungsverbesserung von AKD durch Einsatz hydrophober 

amphoterer Polymere, Wochenbl. für Papierfab. 22, 997-1002. 

Riebeling, U., Jeurissen, H.F.M., Höhr, L. and Schall, N. (1999): 

Wirkungsverbesserung von AKD durch Einsatz hydrophober amphoterer Polymere 

(Teil 2), Wochenbl. für Papierfab. 5, 318-322. 

Roberts, J.C. (1997): A review of advances in internal sizing of paper, The 

Fundamentals of Papermaking, Trans. of the 11th Fund. Res. Symp., Cambridge, 

Pira International, 209-263. 

Roberts, J.C. and Garner, D.N. (1985): The mechanism of alkyl ketene dimer 

sizing of paper. Part 1, Tappi J. 68(4), 118-121. 

Roberts, J.C., Garner, D.N. and Akpabio, U.D. (1985): Neutral sizing and the 

mechanism of alkyl ketene dimer sizing, “Papermaking raw materials” Trans. of 

the eighth Fundamental Res. Symp. held at Oxford, 1985, 2 (3), 815-837. 

Schinzer, D. (1986): Selectives in Lewis Acid Promoted Reactions. Kluwer 

Academic Publishers, Dordrecht, The Netherlands 

Seppänen, R., Tiberg, F. and Valignat, M.-P. (2000): Mechanism of internal 

sizing by alkyl ketene dimers (AKD): The role of the spreading manolayer precourser 

and autophobicity, Nord. Pulp Paper Res. J. 15(5), 452-458. 

Shchukarev, A.V., Mattsson, R. and Ödberg, L. (2003): XPS imaging of surface 

diffusion of alkyl ketene dimer on paper surfaces, Coll. Surf. A: Phys. Chem. Eng. 

Asp. 219(1-3), 35-43. 

Shen, W., Brack, N., Ly, H., Parker, I., Pigram, P.J. and Liesegang, J. (2001a): 

Mechanism of AKD migration studied on glass surfaces, Coll. Surf. A: Phys. 

Chem. Eng. Asp. 176(2-3), 129-137. 

Shen, W. and Parker, I. (2003): An experimental investigation of the redistribution 

behaviour of alkyl ketene dimers and their correponding ketones, Coll. Surf. 

A: Phys. Chem. Eng. Asp. 212(2-3), 197-209. 

Shen, W. and Parker, I.(2003): A study of non-solid behaviour of AKD wax, 

APPITA, 56(6), 442-444. 

Shen, W., Parker, I., Brack, N. and Pigram, P.J. (2001b): A simplified approach 

to understanding the mechanism of AKD sizing, APPITA, 54(4), 352-356. 

Shen, W., Zhang, H. and Ettl, R. (2005): Chemical composition of “AKD vapour” 

and its implication to AKD vapour sizing, Cellulose, 12(6), 641-652. 

Ström, G., Carlsson, G. and Kiær, M. (1992): Bestimmung der Alkylketenedimerverteilung 

in Probeblättern mittels elektron spektroskopie (ESCA), Wochenbl. 

Papierfabr. 120(15), 606-611. 

Thorn, I., Dart, P. ,J. and Main, S. ,D. (1993): The use of cure promotors in alkaline 

sizing, Paper Technol. 34(1), 41-45. 

Walkden, S. ,A. (1991): Alkaline advance helps liquid packaging board meet 

rigorous specifications, Tappi J. 74(4), 103-104. 

Yu, L. and Garnier, G. (2002): The role of vapour deposition during internal 

sizing: A comparative study between ASA and AKD, J. Pulp Paper Sci. 28(10), 

327-331. 

Åvitsland, G., Sterner, M., Wågberg, L. and Ödberg, L. (2006): AKD sizing TCF 

and ECF bleached birch pulp characterized by peroxide edge wicking index, Nord. 

Pulp Paper Res. J. 21(2), 237-244. 

Ödberg, L., Lindström, T., Liedberg, B. and Gustavsson, J. (1987): Evidence 

for β-ketoester formation during the sizing of paper with alkyl ketene dimers, 

Tappi J. 70(4), 135-139. 

Manuscript received October 8, 2007 

Accepted March 16, 2008

Alkyl Ketene Dimer (AKD) sizing – a review

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?